Bitmain's Released Antminer S9, World's First 16nm Miner Ready to Order

[Tupsu]

Between difficulty, electricity and 0.035 BTC a day @ 14TH it will take you a year to break even and three more years to double your money.

It is before 0.035 BTC and after 33 days from now it is only 0.0175 BTC

[1714019866]

1714019866

[hawkfish007]

Oh man I missed the release of this!

 I see everyone still doesn't take into account when you buy something like this u tend to be REPLACING other equipment.
  So to me $2100 is a steal, the equipment I sell I will get back around $1000.
no 220volt so I guess its EVGA 1600watt for $320 another costs to add on.
  And selling off that other equipment drops the electricity used by so much for me. insane. ugh can't believe I missed this!
where are more! lol

Hope these don't fry like usual first batches.

Some folks here think that we are buying S7s with PSUs outright like them. But most people who are buying S9s are replacing their S7s effectively reducing S9 price to $1000 -$1200. Those S7s were in service since September 2015 and already paid themselves off.

[sidehack]

Some of them did. Anyone replacing gear has to make sure and only count the resale value of the S7 once; it can't both contribute to S7 breakeven and S9 initial purchase (in S9 breakeven calculations).
I assume you're replacing 3xS7 with a single S9 to get that reduction?

[hawkfish007]

Some of them did. Anyone replacing gear has to make sure and only count the resale value of the S7 once; it can't both contribute to S7 breakeven and S9 initial purchase (in S9 breakeven calculations).
I assume you're replacing 3xS7 with a single S9 to get that reduction?

Yes, I am considering price of 3 S7s when calculating initial costs of a S9 which reduces it to a more manageable level. My plan is to replace my existing S7s (3 to 1) with S9s and buy more after halving depending on BTC price and difficulty. Please check your messages.

[Puffy23]

Oh man I missed the release of this!

 I see everyone still doesn't take into account when you buy something like this u tend to be REPLACING other equipment.
  So to me $2100 is a steal, the equipment I sell I will get back around $1000.
no 220volt so I guess its EVGA 1600watt for $320 another costs to add on.
  And selling off that other equipment drops the electricity used by so much for me. insane. ugh can't believe I missed this!
where are more! lol

Hope these don't fry like usual first batches.

Some folks here think that we are buying S7s with PSUs outright like them. But most people who are buying S9s are replacing their S7s effectively reducing S9 price to $1000 -$1200. Those S7s were in service since September 2015 and already paid themselves off.

There is no such thing as 'reducing the price'.  You still have to give Bitmain more BTC than the S9 will return.  Whether you purchase that BTC outright or sell off old mining gear you still send the same amount of BTC to Bitmain.  The price is the price and it's higher than the return.  You'd still do better to sell your S7s and hold the coin they make rather than buying an S9.  BTC in > BTC out -> makes a sucker.  

People will jump through whatever hoops they have to to keep from admitting a bad deal...

Oh man I missed the release of this!

 I see everyone still doesn't take into account when you buy something like this u tend to be REPLACING other equipment.
  So to me $2100 is a steal, the equipment I sell I will get back around $1000.
no 220volt so I guess its EVGA 1600watt for $320 another costs to add on.
  And selling off that other equipment drops the electricity used by so much for me. insane. ugh can't believe I missed this!
where are more! lol

Hope these don't fry like usual first batches.

The EVGA SuperNOVA 1600 G2?? Does it have 9 PCI-e connectors?? Will it be able to power an S9?

I believe they have 13 PCI-e.  Maybe 8 6+2-pin and (5) 6-pin.  Something like that.  I've tested an S7 Batch 7 on an extra one I had in the office and it pushed it just fine.  The voltage variation is probably quite a bit more than comparable server PSUs though.

[not.you]

I have 3 S7's from 3 different batches and mining on Kano with greater than 100% luck, none of them have broken even yet, although 2 of them are close if I do not include the PSU.  But I do include the PSU in my ROI calculations since I buy them from newegg with BTC.  I bought brand new EVGA 1600's for the S7's so they were somewhere around 1 BTC each at the time of purchase.  So those two S7's need to generate about 1.2 BTC to break even on the miner itself plus the PSU.  My 3rd S7 is further behind, it needs another half bitcoin to get caught up to the other two, so 1.7 BTC to break even on the miner plus the PSU.  I expect I will probably break even on the miners by the halving but the PSU's may not get paid for for some time.  I could sell the S7's at or close to break even on the miner and upgrade to S9's (just transfer the cost of the PSU's from the S7's to the S9's in my calculations) but I am with Puffy on this, it would cost more BTC then it would generate so totally not worth it at the current price.  If BTC value continues to rise it may get to a point where it is worth it even at $2100 but we got a lot more value to go up before that happens.  I have never purchased BTC with fiat so all of my miners are purchased with BTC I already mined going back a while now.  It makes zero sense for me to buy a miner with BTC I already have that will generate less BTC than it costs.  Until that calculus changes I won't be buying one of these things.

[generalt]

Oh man I missed the release of this!

 I see everyone still doesn't take into account when you buy something like this u tend to be REPLACING other equipment.
  So to me $2100 is a steal, the equipment I sell I will get back around $1000.
no 220volt so I guess its EVGA 1600watt for $320 another costs to add on.
  And selling off that other equipment drops the electricity used by so much for me. insane. ugh can't believe I missed this!
where are more! lol

Hope these don't fry like usual first batches.

I wished there was a 1600w 120v server PSU instead of having to spend $320 on the EVGA.

[philipma1957]

I have 3 S7's from 3 different batches and mining on Kano with greater than 100% luck, none of them have broken even yet, although 2 of them are close if I do not include the PSU.  But I do include the PSU in my ROI calculations since I buy them from newegg with BTC.  I bought brand new EVGA 1600's for the S7's so they were somewhere around 1 BTC each at the time of purchase.  So those two S7's need to generate about 1.2 BTC to break even on the miner itself plus the PSU.  My 3rd S7 is further behind, it needs another half bitcoin to get caught up to the other two, so 1.7 BTC to break even on the miner plus the PSU.  I expect I will probably break even on the miners by the halving but the PSU's may not get paid for for some time.  I could sell the S7's at or close to break even on the miner and upgrade to S9's (just transfer the cost of the PSU's from the S7's to the S9's in my calculations) but I am with Puffy on this, it would cost more BTC then it would generate so totally not worth it at the current price.  If BTC value continues to rise it may get to a point where it is worth it even at $2100 but we got a lot more value to go up before that happens.  I have never purchased BTC with fiat so all of my miners are purchased with BTC I already mined going back a while now.  It makes zero sense for me to buy a miner with BTC I already have that will generate less BTC than it costs.  Until that calculus changes I won't be buying one of these things.

Since you never purchase btc and only mine it that is more. Correct.


I buy btc when I need btc and I sell btc when I need cash.  So. I do not use btc for roi.

I purchased my batch one and batch 2 s-7 s with btc purchased from coinbase.  So I use fiat for my s-7 roi.



@generalalt. yeah I could use that in my friends office.

[Puffy23]

Oh man I missed the release of this!

 I see everyone still doesn't take into account when you buy something like this u tend to be REPLACING other equipment.
  So to me $2100 is a steal, the equipment I sell I will get back around $1000.
no 220volt so I guess its EVGA 1600watt for $320 another costs to add on.
  And selling off that other equipment drops the electricity used by so much for me. insane. ugh can't believe I missed this!
where are more! lol

Hope these don't fry like usual first batches.

I wished there was a 1600w 120v server PSU instead of having to spend $320 on the EVGA.

I know how you feel.  I'd love to run on IBMs 2800s, but I simply don't have the electrical infrastructure to make it work.  I don't have the luxury of starting from scratch and ROI is tough enough without electrical modifications (and what would I do with a ton of 240V when I'm done mining).

That said, ATX PSUs are far too expensive to be a long term solution also.  What I did was grab a few 900W (on 115v) HP-DPS1200s and hard wired them for load sharing.  That effectively gives you an 1,800W server PSU for about $70.  Add the cost of cables and you're still way ahead when compared to EVGA.

Past 750W I don't think it makes sense to buy ATX.  I wish someone would push out a single, or even double, blade version of the high efficiency chips for home mining.

[generalt]

@generalalt. yeah I could use that in my friends office.

Sidehack is working on a dual Dell 750 breakout board with 10 PCIe connectors!
https://bitcointalk.org/index.php?topic=1501428.msg15106357#msg15106357

[dmwardjr]

I have 3 S7's from 3 different batches and mining on Kano with greater than 100% luck, none of them have broken even yet, although 2 of them are close if I do not include the PSU.  But I do include the PSU in my ROI calculations since I buy them from newegg with BTC.  I bought brand new EVGA 1600's for the S7's so they were somewhere around 1 BTC each at the time of purchase.  So those two S7's need to generate about 1.2 BTC to break even on the miner itself plus the PSU.  My 3rd S7 is further behind, it needs another half bitcoin to get caught up to the other two, so 1.7 BTC to break even on the miner plus the PSU.  I expect I will probably break even on the miners by the halving but the PSU's may not get paid for for some time.  I could sell the S7's at or close to break even on the miner and upgrade to S9's (just transfer the cost of the PSU's from the S7's to the S9's in my calculations) but I am with Puffy on this, it would cost more BTC then it would generate so totally not worth it at the current price.  If BTC value continues to rise it may get to a point where it is worth it even at $2100 but we got a lot more value to go up before that happens.  I have never purchased BTC with fiat so all of my miners are purchased with BTC I already mined going back a while now.  It makes zero sense for me to buy a miner with BTC I already have that will generate less BTC than it costs.  Until that calculus changes I won't be buying one of these things.

I, personally, would carry over what remains to be paid towards your PSU's to your next rig (Bitmain S9) in terms of ROI.  You can sell an S7 for $250 plus buyer pays shipping right now.

However, I'm kind of having a hard time myself justifying if it's worth it right now at current price of BTC.  We are at a unique time in terms of our current bullish outlook of bitcoin price compared to the previous two years of bearish markets.

We could probably conclude the buyers of the S9 may do quite well [In terms of monthly profits] during the next three to four months of a rise in bitcoin price that out paces a rise in difficulty.  However, I'm speculating a bitcoin price rise as high as $1,500 the last quarter of 2016.

I'm expecting difficulty to drop down from our current 199,312,067,531 to around 187,000,000,000 by end of July and begin to increase about 5% on average every difficulty change from there until around end of August or September.  I'm expecting at least one manufacturer, whether it's BitFury, KNC, or some other manufacturer, to get next generation rigs up and running soon.  If another manufacturer begins adding next generation rigs to the network, we could see difficulty rises as high as 12% to 17% on average every difficulty change quite easily.  

IF BTC price increases to $1,000 USD, it will have risen 72.4 percent.  If difficulty increased 72.4 percent from 187,000,000,000, that would put difficulty at 322,388,000,000.  When difficulty changes a certain percentage each difficulty change, it is compounded each change.  For example,

If Difficulty was 187 Billion the end of July and increased 5% the first change in August, that would be 196.35 Billion with that 5 percent increase.
2 weeks later (mid August) the difficulty increases another 5% from 196.35 Billion to 206.1675 Billion.
2 weeks later (end of August) the difficulty increases another 5% from 206.1675 Billion to 216.475 Billion.
2 weeks later (mid September) the difficulty increases another 5% from 216.475 Billion to 227.299 Billion.
2 weeks later (end of September) the difficulty increases 15% from 227.299 Billion to 261.394 Billion.  [I have 15% increase at this difficulty change for another manufacturer with next gen rigs coming online.]
2 weeks later (mid October) the difficulty increases 15% from 261.394 Billion to 300.604 Billion.
2 weeks later (end of October) the difficulty increases 11% from 300.604 Billion to 333.670 Billion.  [This puts us above my difficulty prediction of 322,388,000,000 by end of October, 2016.]

An increase in difficulty from 187.000 Billion to 322.388 Billion is an increase of 72.4 percent.  EDIT:  Looking back at the difficulty changes added up, one may think we increased 5% + 5% + 5%+ 5% + 15% + 15% + 11% = 61%.  However, the way it is compounded each increase actually makes this a 72.4 percent increase.

This means a projected rise in difficulty has matched a projected rise in BTC price by end of October, 2016.  Which see, if the price of Bitcoin went up to $1,000 by end of October [72.4 percent increase from current price] and the difficulty is 322.388 Billion [72.4 percent increase from projected lower difficulty in July of 187 Billion].

Here is what projected earnings look like with BTC price @ $1,000 and difficulty at 322,388,000,000:  Half of $655.20 (blocks halved) in a month = $327.60 (0.6552 BTC) before power costs the end of October, 2016.  EDIT:  This assumes you are on a pool with low fees and 100+ percent luck.  I am currently (kano.is).



Since our estimated ROI relies heavily upon our power costs, I'm leaving estimation of ROI up to each miner to calculate.  My estimations could be wrong on the percentage difficulty increases because I'm not sure how many S9's Bitmaintech can manufacture in a month.  I'm also not sure of when another manufacturer will add their next generation rigs to the network and how many they can manufacturer monthly.  

All of this is pure speculation on my part.  I'm offering up my speculation on difficulty for conversation among us all to determine whether it's wise to purchase the S9 or not at current price.

Projected BTC Price Movement on 3D (3-Day) Chart [Meaning each candle is 3-Days in duration]:  https://www.tradingview.com/x/tLfFQO1K//

EDIT:  Also, for those of you trading, I don't see this time being a pump and dump scheme.  That's what the whales want you to think when they fake out a dump, only to find they have created a trap to accumulate more BTC with the intent of pumping it further afterwards.  If this was a pump and dump scheme, they would have started the pump much sooner than what they did.  This time we will have a pump with very little dump and continue pumping again.  2016 will be a bullish year for bitcoin.  2017 through 2019 will be even more bullish than 2016.

[dmwardjr]

not.you asked the following but deleted his question:  "Is that image of the calculator based on 25 coins per block or 12.5?"

I halved the price in my calculation from $655.20 in that month to $327.60 because of blocks reducing from 25 to 12.5.

EDIT:  I'm basically saying your profits with an S9 will look quite nice through October, 2016.  However, don't be surprised if difficulty begins to increase at a rate faster than bitcoin price increase after October, 2016.  Your profits after power costs may reduce from $200+ each month to $100+ each month by December, 2016 IF THE PRICE OF BITCOIN DOES NOT INCREASE TO THE PROJECTED $1,500 I've speculated in my 3-Day price chart by end of 2016.  However, if bitcoin price does increase to $1,500 by end of 2016, you can expect the percentage rate of difficulty to increase at a similar percentage rate with bitcoin price until the end of 2016 and continue to enjoy $200+ profits per S9 monthly.  However, I can see the difficulty increase exponentially in 2017 if the price of bitcoin increased to $1,500.

[CarrollFilms]

14Th/s sounds like a nice big number. I threw it into a mining calculator and even with free electricity and the halving reward put in it would take 200+ days. But that is with a fixed a difficulty.

I feel sorry for the folks that will end up buying one of these at the price they're currently at.

I'm hoping, either, higher hashing power machines come out, or they are made and sold cheaper.

Keep in mind it could change with BTC value.  Say you put in the 2kish and mine if there keeps being surges in BTC price that will also help for longer mining.  Leading up to having btc value has been nice. 

After having some miners might come offline to due to efficiency, it's all speculation.  S9's with amazing efficiency they will be able to run after having no doubt.  So again difficulty I think is hard to say after having.   There are some circumstances that would make it a good investment, but mining never has been a sure thing.

It's hard to predict the amount of hashrate going away because of inefficient miners. Nor any price rise because of the halving. Which makes it quite risky to buy a miner where you got no guarantee  of how much it will let you earn you.

Suppose the price goes back to 400 after the halving, then you are screwed...

That's not going to happen unless there is another exchange scam or China decides to dump it all

[dmwardjr]

That's not going to happen unless there is another exchange scam or China decides to dump it all

I would have to agree.  I believe the day of seeing sub $450 USD for a bitcoin and sub $4.50 for LTC is history.

[sidehack]

So if coin increases to $1500, should I buy 1.4BTC right now for about $800 and cash it out for $2100 at the end of the year, or buy 3.6BTC right now for about $2100 and mine around 1.4 BTC by the end of the year to cash out for $2100?

[Biodom]

So if coin increases to $1500, should I buy 1.4BTC right now for about $800 and cash it out for $2100 at the end of the year, or buy 3.6BTC right now for about $2100 and mine around 1.4 BTC by the end of the year to cash out for $2100?

<put your name here> in the alt Universe already bought those 1.4 BTC last year for $400  

[chiguireitor]

Good point. ASIC efficiency is probably about to plateau for a while. One consideration to add to that is S2 vs S1 - a more chip-dense machine with the same ASIC about half the per-hash power (for the current generation, probably about 0.6J/GH machine-level) will have a higher initial cost but a substantially longer viability. In that way I guess Bitmain has done hackers a favor going with a bucked string topology on later S7 and S9. For a regular joe it's actually worse than unregulated string because you get worse efficiency but if it can be hacked to adjustable voltage that potentially doubles (or more) the miner's effective lifetime.

-removed for privacy-

...

Sidehack is working on a dual Dell 750 breakout board with 10 PCIe connectors!
https://bitcointalk.org/index.php?topic=1501428.msg15106357#msg15106357

-removed for privacy-

[fr4nkthetank]

I like these speculations.  Another way to view it this:  If you have s7's that you want to sell, sell now or wait ?  We know miners hit a "floor" value at some point.  So is selling an s7 for 250$ reasonable right now ?  How much will it be worth after halving ?  In october ?  100$?  Maybe its better to keep it ?  I too feel (and I have said before) we are due for a massive pump, and the last one took at least a quarter to materialize, this one will be longer.  Mayyyybe, just maybe, the smart move is to keep the s7's until they hit "floor" value, and just keep that sweet btc, and dont give it to Bitmain.  Then at some point switch em for the s9's or actually skip this generation.  Its a tough call.  Like I feel btc will really double or triple, and i'd buy a ton of it, but unfortunately money doesnt grow on trees.

[philipma1957]

So if coin increases to $1500, should I buy 1.4BTC right now for about $800 and cash it out for $2100 at the end of the year, or buy 3.6BTC right now for about $2100 and mine around 1.4 BTC by the end of the year to cash out for $2100?

<put your name here> in the alt Universe already bought those 1.4 BTC last year for $400  

I made more money in two months with eth coin then I made with btc in the last seven months.

And in invested less in eth coin then I did btc.

[sidehack]

I've got about half a dozen S7 running in hosting on dual Dell 750s. The difference with a new batch will be the option for 6-pin cabling, which people seem to like because, though far less versatile, it's quick and clean. I'd probably be selling these boards out the butt if I had any for the last six months, but things have been tight around the shop between Compacs and DPS8/12 boards and people waiting a month and a half to pay their bills. Money definitely does not grow on trees around here.

A 208V-fed DPS1200 will run a 135-chip S7 on stock settings. Runs hot, but it runs.

Good schematics would be nice. Heck of a lot easier than reverse-engineering, and cheaper. Tomorrow's task will be hacking away at Avalon stuff since they didn't provide an A3218 datasheet. After that will be hacking away at an S7LN board. Making things work and making things work better are always fun.